@grida/ai-models 1.20261113.0 → 2.20261115.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/PROVENANCE.md CHANGED
@@ -1892,3 +1892,21 @@ medium 560 and ultra_high 2,240. This maximum comes from the allocation table, n
1892
1892
  the model page's default resolution. Image MIME types are sourced separately.
1893
1893
  https://aistudio.google.com/docs/media-resolution
1894
1894
  https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/gemini/4-hydrogen
1895
+
1896
+ ## Grida Gemini 4 Hydrogen discovery — 2026-11-13
1897
+
1898
+ [Grida model discovery](https://gateway.grida.ai/v1/models) and the
1899
+ [public API reference](https://gateway.grida.ai/llms.txt) list
1900
+ `google/gemini-4-hydrogen` through Chat Completions and Responses with text and
1901
+ image input, tools, reasoning and structured output. The published combined
1902
+ input/output window is 1,000,000 tokens; output includes thinking, defaults to
1903
+ 32,768 tokens and is capped at 65,536 with or without streaming.
1904
+
1905
+ The public binding records regular USD-per-token components of $0.0000015
1906
+ ordinary input, $0.000000075 cached input and $0.0000075 output, each with
1907
+ one-token quantities and no additional quantity rounding. Explicit effort accepts
1908
+ low, medium, high and xhigh. Omitting effort leaves the upstream setting
1909
+ unspecified; shared effort names do not promise a particular native thinking
1910
+ level. Sampling controls, explicit caches, grounding, file/audio/video input,
1911
+ background generation and continuation are outside this service contract.
1912
+ Native model facts and other provider bindings remain unchanged.
package/README.md CHANGED
@@ -5,7 +5,7 @@ provider bindings, and published prices, with the labs that make them and their
5
5
  logos. Every fact cites a first-party source.
6
6
 
7
7
  This package is the data, published unchanged from release
8
- [`v1.20261113.0`](https://github.com/gridaco/models/releases/tag/v1.20261113.0) of [gridaco/models](https://github.com/gridaco/models). It contains no code.
8
+ [`v2.20261115.0`](https://github.com/gridaco/models/releases/tag/v2.20261115.0) of [gridaco/models](https://github.com/gridaco/models). It contains no code.
9
9
 
10
10
  | file | contents |
11
11
  | --- | --- |
@@ -15,6 +15,7 @@ This package is the data, published unchanged from release
15
15
  | `schemas/` | JSON Schemas for `models.json` and `labs.json` |
16
16
  | `manifest.json` | Version, source commit, content digest, and the SHA-256 of every file |
17
17
  | `PROVENANCE.md` | Source notes that do not fit in JSON |
18
+ | `SCHEMA_CHANGELOG.md` | Authored schema changes, compatibility and migration notes |
18
19
 
19
20
  ```js
20
21
  import models from "@grida/ai-models/models.json" with { type: "json" };
@@ -24,13 +25,13 @@ import labs from "@grida/ai-models/labs.json" with { type: "json" };
24
25
  From a CDN, without installing:
25
26
 
26
27
  ```
27
- https://cdn.jsdelivr.net/npm/@grida/ai-models@1/models.json
28
- https://cdn.jsdelivr.net/npm/@grida/ai-models@1.20261113.0/models.json
28
+ https://cdn.jsdelivr.net/npm/@grida/ai-models@2/models.json
29
+ https://cdn.jsdelivr.net/npm/@grida/ai-models@2.20261115.0/models.json
29
30
  ```
30
31
 
31
32
  ## Versions
32
33
 
33
- Versions read `<schema>.<YYYYMMDD>.<n>`. The major is the schema version, so `^1` follows
34
+ Versions read `<schema>.<YYYYMMDD>.<n>`. The major is the schema version, so `^2` follows
34
35
  every snapshot without crossing a breaking change. A new version is published whenever the
35
36
  catalogue's content changes.
36
37
 
@@ -0,0 +1,78 @@
1
+ # Schema changelog
2
+
3
+ Authored history of the catalogue's public contract: fields, validation rules and
4
+ their meanings. Model additions, provider coverage and price changes belong in
5
+ the generated [snapshot release notes](https://github.com/gridaco/models/releases).
6
+
7
+ Entries are grouped by schema major, newest first. Entry dates identify when a
8
+ change was documented; tagged releases establish when a snapshot was published.
9
+ An entry linked to an open PR does not mean the change has been released. See
10
+ [Maintaining this log](#maintaining-this-log) for contributor requirements.
11
+
12
+ ## Schema v2
13
+
14
+ ### 2026-11-15 — Decision offerings may omit an underlying model
15
+
16
+ Reference: [#44](https://github.com/gridaco/models/pull/44).
17
+
18
+ **Change:** `decision_card.underlying_model` is optional. It continues to record
19
+ the canonical identity of a documented underlying model when present, with the
20
+ existing nonempty-string constraint. Omission means no relationship is recorded;
21
+ it does not infer a parent, a self-reference or inherited chat capabilities.
22
+
23
+ The validation change is:
24
+
25
+ ```diff
26
+ - "required": ["vendor", "underlying_model", "providers"]
27
+ + "required": ["vendor", "providers"]
28
+ ```
29
+
30
+ **Compatibility:** breaking for v1 readers that require this relationship. The
31
+ new validator accepts cards with the relationship retained, but a newly emitted
32
+ card without it fails the old v1 validator. The Models `$schema`, both schema
33
+ `$id`s and release/package major move to v2. Labs retain their record shape and
34
+ follow the same release major.
35
+
36
+ `decision.models` remains semantic grouping, not a shared execution protocol.
37
+ Provider request/result contracts, uncertainty meanings, usage and prices remain
38
+ independent. No new primitive set, model card or published rate is introduced.
39
+
40
+ **Migration:** update the reader/validator before selecting a v2 snapshot and
41
+ handle an absent relationship. Existing v1 pins remain on v1; GitHub `latest`
42
+ moves across majors, so moving v1 asset URLs stop resolving after the first v2
43
+ release. See the [v1-to-v2 migration instructions](https://github.com/gridaco/models/blob/5fd023df088af38a135246c8e476f60d7c30cbc6/README.md#migrating-from-schema-version-1)
44
+ for schema identifiers, asset names and pin choices.
45
+
46
+ ## Schema v1
47
+
48
+ Version 1 predates this authored log. Its contract history remains in the
49
+ [tagged snapshots and release notes](https://github.com/gridaco/models/releases)
50
+ and [schema Git history](https://github.com/gridaco/models/commits/main/catalogue/schemas).
51
+ This log starts with the v2 transition rather than reconstructing an incomplete
52
+ history of v1 changes.
53
+
54
+ ## Maintaining this log
55
+
56
+ Update this file in the same PR as a public contract change, including compatible
57
+ extensions. This covers added/removed fields, requiredness, types, enum values,
58
+ constraints and field meanings, plus consumer-visible validation or generation
59
+ behavior implemented outside the JSON Schemas. A new model fitting the existing
60
+ contract needs no entry.
61
+
62
+ Keep each entry short: state the field/rule and before/after behavior, whether
63
+ the change is breaking, the consumer action needed (or why none is needed), and
64
+ the PR link. Add dated entries under the relevant schema major; changing snapshot
65
+ dates or counters does not create another entry. Follow the catalogue's existing
66
+ [versioning rules](https://github.com/gridaco/models/blob/main/catalogue/README.md#rules)
67
+ for a breaking change.
68
+
69
+ Exclude model facts, model/provider additions, rates, provenance-only corrections,
70
+ tooling-only refactors and formatting-only changes. Preserve published history;
71
+ make corrections explicit instead of silently rewriting an earlier contract's
72
+ meaning.
73
+
74
+ The PR check compares JSON content under `catalogue/schemas/` and requires a
75
+ non-whitespace changelog update when that content changes, including schema
76
+ additions or removals. It ignores JSON formatting and object-key order. This
77
+ check cannot judge prose accuracy or detect every behavioral change outside the
78
+ schemas: contributors and reviewers must check those changes against this log.
package/manifest.json CHANGED
@@ -1,23 +1,27 @@
1
1
  {
2
2
  "name": "grida-models",
3
- "version": "v1.20261113.0",
4
- "schema_version": 1,
5
- "commit": "22659ac127a10b56a87bd541020c7ee8e1f75091",
6
- "created": "2026-11-13T09:32:26Z",
7
- "source": "https://github.com/gridaco/models/tree/22659ac127a10b56a87bd541020c7ee8e1f75091/catalogue",
8
- "digest": "sha256:4cc3236c84546e3fd114d6b2b4143d281c0238d859dde46bf680f7c74112f24e",
3
+ "version": "v2.20261115.0",
4
+ "schema_version": 2,
5
+ "commit": "26cd2de507d29961546dabe66d83fbb2063d0bc0",
6
+ "created": "2026-11-15T10:57:18Z",
7
+ "source": "https://github.com/gridaco/models/tree/26cd2de507d29961546dabe66d83fbb2063d0bc0/catalogue",
8
+ "digest": "sha256:cf33026409454b86b5da940a849f41f6a1509dd76ba1918ee5b29210eaf181d0",
9
9
  "counts": {
10
10
  "cards": 106,
11
11
  "labs": 18
12
12
  },
13
13
  "files": {
14
14
  "PROVENANCE.md": {
15
- "sha256": "4be6aa57f189a5072de0396227f1f68b2b1ee4029e2cc96aa44e71c1d99a7f03",
16
- "bytes": 101464
15
+ "sha256": "b691da0cb87742ff9667a9bb6362f59655c24e14677555817a7e30b9c14c0fd8",
16
+ "bytes": 102548
17
17
  },
18
18
  "README.md": {
19
- "sha256": "c4251f78423708a1c6cfbe0cfa4c208fb7cecbfc14ddd512bb46c5671abeae7a",
20
- "bytes": 28706
19
+ "sha256": "02a012d4a62236c014e5f655e31d4e934981520788bde96788d1db3ba3eb9322",
20
+ "bytes": 31381
21
+ },
22
+ "SCHEMA_CHANGELOG.md": {
23
+ "sha256": "4ce477fc504b676055aa4d713a75879327c1a001efa55c4cfacd886682aec17c",
24
+ "bytes": 4043
21
25
  },
22
26
  "labs.json": {
23
27
  "sha256": "b3267fae072fb4b6083f464117a9d5109b5b1237211b8541ecb083b86fd26c53",
@@ -92,16 +96,16 @@
92
96
  "bytes": 330
93
97
  },
94
98
  "models.json": {
95
- "sha256": "156ce2cd0a3210525a7ccc58fd847a9c53c2988f3d7f4dec3960f89cbea2f94f",
96
- "bytes": 1121561
99
+ "sha256": "c01b4ed634e45e5e45b94ec05fa957f73e99e2f866b7028ff79217c29dd692ec",
100
+ "bytes": 1123858
97
101
  },
98
102
  "schemas/labs.schema.json": {
99
- "sha256": "9ddaedd56826bebdbe72b84f76cfc86df0b5c5902ba9f0259c2ea2cbb9327f99",
103
+ "sha256": "a720f50f9ee5c3ddd2fccf50e38ddecec686896b768a52cc987df52668171f07",
100
104
  "bytes": 2469
101
105
  },
102
106
  "schemas/models.schema.json": {
103
- "sha256": "f76f411371e0e22bb40c72e2c33f68b091dabcad81741d6e60e07f9af0395aa1",
104
- "bytes": 40220
107
+ "sha256": "d095009e7d362f70d37116a90a48f9b44fd252d4c578983c79076fc6eba212c9",
108
+ "bytes": 40443
105
109
  }
106
110
  }
107
111
  }
package/models.json CHANGED
@@ -1,5 +1,5 @@
1
1
  {
2
- "$schema": "https://grida.co/schemas/ai-model-facts-v1.json",
2
+ "$schema": "https://grida.co/schemas/ai-model-facts-v2.json",
3
3
  "text.catalog": {
4
4
  "anthropic/claude-fable-5": {
5
5
  "id": "anthropic/claude-fable-5",
@@ -3773,6 +3773,69 @@
3773
3773
  "output": "0.0000075",
3774
3774
  "input_cache_read": "0.000000075"
3775
3775
  }
3776
+ },
3777
+ "grida": {
3778
+ "provider": "grida",
3779
+ "id": "google/gemini-4-hydrogen",
3780
+ "url": "https://gateway.grida.ai/v1/models",
3781
+ "endpoints": [
3782
+ {
3783
+ "id": "google/gemini-4-hydrogen",
3784
+ "input": "text",
3785
+ "source_url": "https://gateway.grida.ai/v1/models",
3786
+ "published_pricing": {
3787
+ "source_url": "https://gateway.grida.ai/v1/models",
3788
+ "currency": "USD",
3789
+ "components": [
3790
+ {
3791
+ "id": "input_tokens",
3792
+ "meter": "input_tokens",
3793
+ "unit": "token",
3794
+ "per_quantity": "1",
3795
+ "when": {},
3796
+ "quantity_rounding": {
3797
+ "mode": "none"
3798
+ },
3799
+ "usd_per_unit": "0.0000015"
3800
+ },
3801
+ {
3802
+ "id": "cached_input_tokens",
3803
+ "meter": "cached_input_tokens",
3804
+ "unit": "token",
3805
+ "per_quantity": "1",
3806
+ "when": {},
3807
+ "quantity_rounding": {
3808
+ "mode": "none"
3809
+ },
3810
+ "usd_per_unit": "0.000000075"
3811
+ },
3812
+ {
3813
+ "id": "output_tokens",
3814
+ "meter": "output_tokens",
3815
+ "unit": "token",
3816
+ "per_quantity": "1",
3817
+ "when": {},
3818
+ "quantity_rounding": {
3819
+ "mode": "none"
3820
+ },
3821
+ "usd_per_unit": "0.0000075"
3822
+ }
3823
+ ]
3824
+ }
3825
+ }
3826
+ ],
3827
+ "capabilities": {
3828
+ "source_url": "https://gateway.grida.ai/v1/models",
3829
+ "image_input": true,
3830
+ "tools": true,
3831
+ "reasoning": true,
3832
+ "structured_output": true
3833
+ },
3834
+ "limits": {
3835
+ "source_url": "https://gateway.grida.ai/v1/models",
3836
+ "context_tokens": 1000000,
3837
+ "max_output_tokens": 65536
3838
+ }
3776
3839
  }
3777
3840
  },
3778
3841
  "provider_coverage": {
@@ -3792,6 +3855,10 @@
3792
3855
  "disposition": "excluded",
3793
3856
  "source_url": "https://fal.ai/models",
3794
3857
  "reason": "The provider coverage policy includes fal for media models only; no text API compatibility is claimed."
3858
+ },
3859
+ "grida": {
3860
+ "disposition": "listed",
3861
+ "source_url": "https://gateway.grida.ai/v1/models"
3795
3862
  }
3796
3863
  }
3797
3864
  },
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@grida/ai-models",
3
- "version": "1.20261113.0",
3
+ "version": "2.20261115.0",
4
4
  "description": "A hand-curated, source-backed catalogue of AI models, their labs, and published prices.",
5
5
  "keywords": [
6
6
  "ai",
@@ -20,6 +20,7 @@
20
20
  "license": "MIT",
21
21
  "files": [
22
22
  "PROVENANCE.md",
23
+ "SCHEMA_CHANGELOG.md",
23
24
  "labs.json",
24
25
  "logos",
25
26
  "manifest.json",
@@ -1,7 +1,7 @@
1
1
  {
2
2
  "$schema": "https://json-schema.org/draft/2020-12/schema",
3
- "$id": "https://grida.co/schemas/ai-model-labs-v1.json",
4
- "title": "Model labs, version 1",
3
+ "$id": "https://grida.co/schemas/ai-model-labs-v2.json",
4
+ "title": "Model labs, version 2",
5
5
  "description": "The labs that make catalogued models, keyed by lab ID. A lab ID is the `vendor` value on models.json cards.",
6
6
  "type": "object",
7
7
  "minProperties": 1,
@@ -1,7 +1,7 @@
1
1
  {
2
2
  "$schema": "https://json-schema.org/draft/2020-12/schema",
3
- "$id": "https://grida.co/schemas/ai-model-facts-v1.json",
4
- "title": "Canonical model facts, authoring envelope version 1",
3
+ "$id": "https://grida.co/schemas/ai-model-facts-v2.json",
4
+ "title": "Canonical model facts, authoring envelope version 2",
5
5
  "description": "Structural interchange schema with sourced, additive model and serving-provider facts.",
6
6
  "type": "object",
7
7
  "additionalProperties": false,
@@ -22,7 +22,7 @@
22
22
  "properties": {
23
23
  "$schema": {
24
24
  "description": "This file's schema: the `$id` of models.schema.json. Generated files always carry it.",
25
- "const": "https://grida.co/schemas/ai-model-facts-v1.json"
25
+ "const": "https://grida.co/schemas/ai-model-facts-v2.json"
26
26
  },
27
27
  "text.catalog": {
28
28
  "type": "object",
@@ -209,6 +209,7 @@
209
209
  }
210
210
  },
211
211
  "decision.models": {
212
+ "description": "Semantic grouping of decision offerings, not a shared execution protocol. Each provider endpoint retains its own request, result and accounting contract.",
212
213
  "type": "object",
213
214
  "minProperties": 1,
214
215
  "additionalProperties": {
@@ -1339,13 +1340,13 @@
1339
1340
  "$ref": "#/$defs/card"
1340
1341
  }
1341
1342
  ],
1342
- "required": ["vendor", "underlying_model", "providers"],
1343
+ "required": ["vendor", "providers"],
1343
1344
  "unevaluatedProperties": false,
1344
1345
  "properties": {
1345
1346
  "underlying_model": {
1346
1347
  "type": "string",
1347
1348
  "minLength": 1,
1348
- "description": "Canonical catalogue identity of the underlying model; this offering has its own interface and rates."
1349
+ "description": "Canonical catalogue identity of a documented underlying model, when known. Omission means no relationship is recorded; this offering has its own interface and rates."
1349
1350
  },
1350
1351
  "decision_contract": {
1351
1352
  "$ref": "#/$defs/decision_contract"